- Email: [email protected]
Impact of neurohormonal blockade on association between body mass index and mortality
International Journal of Cardiology 119 (2007) 33 – 40 www.elsevier.com/locate/ijcard
Impact of neurohormonal blockade on association between body mass index and mortality Linn M.A. Kennedy a,⁎, Stefan D. Anker b,c , John Kjekshus d , Tom J. Cook e , Ronnie Willenheimer a a
c
Department of Cardiology, Malmö University Hospital, Lund University, Malmö, Sweden b Department of Clinical Cardiology, National Heart and Lung Institute, London, UK Division of Applied Cachexia Research, Department of Cardiology, Charité Campus Virchow Klinikum, Berlin, Germany d Department of Cardiology, Stavanger University Hospital, Oslo, Norway e Merck and Co., Whitehouse Station, NJ, USA Received 13 February 2006; accepted 19 June 2006 Available online 17 October 2006
Abstract Background: The prognostic impact of body mass index (BMI) in patients following acute myocardial infarction (AMI) may be altered by neurohormonal blockade. Methods: The impact of neurohormonal blockade on the association between BMI and mortality was examined in 5548 patients following AMI (CONSENSUS II), 50% receiving enalapril and 7% beta-blockade, and in 4367 patients with coronary artery disease (CAD) (4S), 79% with prior AMI, 12% receiving ACEi and 67% beta-blockade. Median follow-up was 0.4 and 5.2 years, respectively. Patients were categorized into 4 BMI groups: Underweight, b 22.00; normal-weight, 22.00–24.99; overweight, 25.00–29.99; obese, ≥ 30.00 kg/m2. Multivariable analysis adjusted for demographics, patient history, physical examination, biochemistry and medication. Results: CONSENSUS II: Overall, adjusted mortality (n = 301) risk was similar across BMI groups. Comparing overweight with normalweight patients, the hazard ratios (HRs) for mortality differed significantly (P = 0.028) between patients randomized to placebo (HR 1.41) and enalapril (HR 0.75). 4S: Overall, adjusted mortality (n = 421) risk was similar for normal-weight, overweight and obese patients. In a time-dependent analysis for drug use, comparing obese with normal-weight patients, the HRs for mortality differed significantly (P = 0.047) between patients without (HR 1.86) and those with (HR 0.97) neurohormonal blockade. Conclusion: In patients after AMI or with CAD, high BMI was associated with increased mortality risk among patients not receiving neurohormonal blockade, but with decreased or neutral mortality risk among those receiving neurohormonal blockade. Tests for interaction indicate that neurohormonal blockade may attenuate the relationship between high BMI and increased mortality risk. Neurohormonal blockade may thus partly explain the so-called obesity paradox. © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Acute myocardial infarction; Coronary artery disease; Body mass index; Neurohormonal blockade; Mortality
1. Introduction Obesity is a serious independent risk factor for coronary artery disease and premature death, and is related to cardiovascular risk factors such as hypertension, dyslipidemia and insulin resistance [1,2]. However, the prognostic importance ⁎ Corresponding author. Tel.: +46 40 331000; fax: +46 40 336209. E-mail address: [email protected] (L.M.A. Kennedy). 0167-5273/$ - see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ijcard.2006.06.064
of overweight and obesity in patients with coronary artery disease and following acute myocardial infarction is not yet fully understood. Studies have shown diverse and even contradictory results; high body mass index (BMI) has had a negative, neutral or even positive effect on survival [3–16]. In patients with coronary artery disease undergoing percutaneous coronary intervention, overweight and obesity have been associated with a better prognosis, a finding sometimes referred to as “the obesity paradox”[7].
34
L.M.A. Kennedy et al. / International Journal of Cardiology 119 (2007) 33–40
There is evidence that local adipose tissue and systemic renin–angiotensin system activation, as well as obesityrelated increased sympathetic activity may play an important role in the pathophysiology of obesity-related cardiovascular disorders [17,18]. Hence, it may be of benefit to block the renin–angiotensin system and the sympathetic nervous system in overweight and obese patients. In a previous analysis of the OPTIMAAL trial we found that, in 5477 patients with complicated acute myocardial infarction, overweight and obese patients had similar adjusted mortality risk compared to normal-weight patients, during a follow-up of 2.8 years [19]. These findings might possibly be due to the effects of renin– angiotensin system inhibition by angiotensin-converting enzyme inhibitors or angiotensin-receptor blockers, and/or inhibition of the sympathetic nervous system by betablockade; all patients received an angiotensin-converting enzyme inhibitor or an angiotensin-receptor blockers and 75% were given beta-blockade in the OPTIMAAL trial [19]. Therefore, we hypothesized that overweight/obesity is associated with increased mortality risk in patients not receiving neurohormonal blockade – i.e. angiotensinconverting enzyme inhibitors, angiotensin-receptor blockers or betablockers – but not in patients receiving such treatment. In order to test our hypothesis we evaluated the impact of treatment with neurohormonal blockade on the relationship between BMI and prognosis in patients after acute myocardial infarction or with stable coronary artery disease, included in the CONSENSUS II [20] and 4S trials [21]. 2. Methods 2.1. Patients and trial design The 6090 patients in the CONSENSUS II trial had acute myocardial infarction and were randomized to double-blind treatment with enalapril or placebo [20]. Beta-blockers were optional. The primary endpoint was all-cause mortality. The 4444 patients included in the 4S trial had angina pectoris or previous acute myocardial infarction (79%) and received randomized double-blind treatment with simvastatin or placebo [21]. Angiotensin-converting enzyme inhibitors and betablockers were optional. The primary endpoint was allcause mortality.
(22.00–24.99 kg/m2), overweight (25.00–29.99 kg/m2), obese (N/= 30 kg/m2). In multivariable analyses, adjustment was made for several baseline variables comprising demographics, patient history, medication, physical examination and biochemical analyses. 2.3. Statistics The value of baseline BMI group to predict mortality was tested in uni- and multivariable Cox Proportional Hazard models. In a second step, we tested the impact of treatment with neurohormonal blockade on the association between baseline BMI group and mortality. Due to statistical power issues, these analyses were performed with the overweight and obese groups separately as well as combined. Formal tests for interaction depending on treatment with neurohormonal blockade were performed. In CONSENSUS II, in-hospital use of betablockers and angiotensin-converting enzyme inhibitors was considered use of that treatment. Due to few events in the analyses when dividing the material, test for interaction was only performed in univariable model in CONSENSUS II. In 4S treatment with betablockers and angiotensin-converting enzyme inhibitors was used as a time dependent variable. In this analysis, patients were considered to be on treatment if they had been so for at least 14 days, and to be off treatment if they had not received any such treatment within 14 days. Test for interaction was performed in a multivariable model. Baseline covariables were screened for possible inclusion in a multivariable model. Hazard ratios (HR) and 95% confidence intervals (CI) were estimated for the effect of BMI group relative to the normal BMI group. Covariables found to be significantly (P b 0.05) associated with mortality in univariable analysis were tested for internal correlation, using Spearman rank correlation test for categorical variables and linear regression analysis for continuous variables. In all analyses, R N 0.3 denoted significant internal correlation, and only the variable with the greatest association with mortality was qualified for further analysis. Variables remaining after this procedure were included in the corresponding multivariable model. The covariable selection method resulted in the inclusion into the CONSENSUS II multivariable model of baseline loop diuretic use, systolic blood pressure, heart rate, age, creati-
2.2. Design of the present analysis
Table 1 Baseline BMI by baseline BMI group
In post-hoc analysis we assessed the impact of treatment with an angiotensin-converting enzyme inhibitor and/or a betablocker on the association between baseline BMI group and mortality (no patient received an angiotensin-receptor blockers in CONSENSUS II or 4S). Prior to data analysis, patients were categorized into 4 groups according to baseline BMI: underweight (b22.00 kg/m2), normal-weight
BMI group
Underweight Normal Overweight Obese BMI, body mass index.
CONSENSUS II
4S
N
%
N
%
640 1624 2572 712
11.5 29.3 46.4 12.8
364 1474 2060 469
8.3 33.8 47.2 10.7
L.M.A. Kennedy et al. / International Journal of Cardiology 119 (2007) 33–40
35
Table 2a CONSENSUS II: Baseline characteristics by baseline BMI group Baseline characteristic
Baseline BMI group Underweight
Normal
N = 640
Age (years) Pulse rate (beats/min) Potassium (mmol/L) Sodium (mmol/L) Serum creatinine (mmol/L) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg)
Female gender History of AMI History of hypertension History of diabetes History of CHF Baseline statin use Baseline aspirin use Baseline beta blocker use Baseline diuretic use Baseline warfarin use
Overweight
N = 1624
Obese
N = 2572
N = 712
Mean
S.D.
Mean
S.D.
Mean
S.D.
Mean
S.D.
67.93 76.51 4.05 139.42 94.28 130.73 77.37
11.34 16.58 0.44 3.19 25.28 19.50 11.97
66.64 74.14 4.04 139.60 95.70 131.93 79.24
11.04 14.58 0.44 3.19 23.04 19.13 11.89
64.36 74.27 4.04 139.67 96.82 133.78 80.49
11.01 15.00 0.43 3.09 22.47 19.98 12.14
63.49 77.58 4.06 139.65 95.76 138.09 82.91
11.29 16.60 0.44 3.18 26.58 21.30 12.64
n
%
n
%
n
%
n
%
247 141 128 57 41 4 11 83 267 61
38.59 22.03 20.00 8.91 6.41 0.63 1.72 12.97 41.72 9.53
378 381 351 141 95 17 30 230 614 171
23.28 23.46 21.61 8.68 5.85 1.05 1.85 14.16 37.81 10.53
590 591 726 286 121 15 35 399 962 244
22.94 22.98 28.23 11.12 4.70 0.58 1.36 15.51 37.40 9.49
257 186 299 133 76 6 9 120 340 42
36.10 26.12 41.99 18.68 10.67 0.84 1.24 16.85 47.75 5.90
nine, potassium, sodium and total lactate dehydrogenase. In 4S, the multivariable model covariable selection resulted in the inclusion into the model of age, gender, smoking status, hemoglobin, history of diabetes, hypertension and claudication, randomized treatment with simvastatin/placebo, and inclusion into the 4S trial due to acute myocardial infarction or not.
3. Results In both studies, around half of the patients were overweight, one quarter were normal-weight and a little more than 10% were obese (Table 1). Baseline characteristics of the patients included in the respective trials are seen in Table 2a (CONSENSUS II) and Table 2b (4S). In
Table 2b 4S: Baseline characteristics by baseline BMI group Baseline characteristic
Age (years) Glucose (mmol/L) HDL (mmol/L) Hemoglobin (mmol/L) LDL cholesterol (mmol/L) Pulse rate (beats/min) Serum creatinine (mmol/L) Systolic blood pressure (mm Hg) Triglycerides (mmol/L)
Female gender Current smoker History of diabetes History of AMI Baseline aspirin Baseline beta blocker Baseline diuretic Baseline ACEi and/or beta blocker
Baseline BMI group Underweight
Normal
Overweight
Obese
N = 364
N = 1474
N = 2060
N = 469
Mean
S.D.
59.2 99.5 1.35 85.7 4.77 64.5 87.1 135.4 1.26
7.0 20.3 0.34 65.7 0.67 10.7 19.8 20.0 0.43
Mean
n
%
n
%
118 131 9 296 123 181 11 185
32.4 36.0 2.5 81.3 33.8 49.7 3.0 50.8
262 414 41 1179 570 776 81 793
17.8 28.1 2.8 80.0 38.7 52.6 5.5 53.8
59.0 102.3 1.21 86.1 4.94 64.0 90.6 136.5 1.41
S.D. 7.0 21.2 0.30 67.9 0.67 10.1 21.2 19.3 0.50
Mean
S.D.
Mean
S.D.
7.12 26.98 0.28 0.69 0.65 9.92 15.06 19.17 0.50
57.86 111.44 1.13 9.34 4.85 64.84 92.24 144.20 1.69
7.26 30.13 0.25 0.72 0.64 9.71 15.03 20.43 0.50
n
%
n
%
308 457 109 1625 747 1237 122 1260
14.95 22.18 5.29 78.88 36.26 60.05 5.92 61.17
122 98 42 373 164 311 61 318
26.01 20.90 8.96 79.53 34.97 66.31 13.01 67.80
58.33 106.22 1.15 9.28 4.86 63.36 93.47 139.81 1.56
36
L.M.A. Kennedy et al. / International Journal of Cardiology 119 (2007) 33–40
Table 3 CONSENSUS II: Neurohormonal blockade by baseline BMI group Neurohormonal blockade
ACEi Beta-blocker ACEi and betablocker Neither ACEi nor betablocker
Underweight
Normal
Overweight
Obese
n
%
n
%
n
%
n
%
295 43 40 262
46.09 6.72 6.25 40.94
697 114 116 697
42.92 7.02 7.14 42.92
1036 202 197 1137
40.28 7.85 7.66 44.21
317 56 64 275
44.52 7.87 8.99 38.62
4S: Neurohormonal blockade use at least 14 days during study by baseline BMI group Neurohormonal blockade
ACEi Betablocker ACEi and/or betablocker
Underweight
Normal
Overweight
Obese
n
%
n
%
n
%
n
%
28 209 220
7.7 57.4 60.4
148 918 970
10.0 62.3 65.8
252 1432 1501
12.2 69.5 72.9
85 353 371
18.12 75.27 79.10
ACEi, angiotensin-converting enzyme inhibitor.
CONSENSUS II, the follow-up period had a median length of 0.4 years (139 days, range 2–181 days) and in 4S it was 5.2 years (range 11–2267 days). During follow-up, there were 301 deaths in CONSENSUS II and 421 deaths in 4S. 3.1. CONSENSUS II The use of neurohormonal blockade was similar across the baseline BMI groups (Table 3). In univariable analysis including all patients, compared to normal-weight patients, the overall mortality risk was significantly higher in underweight patients (HR 1.64, 95% CI 1.16–2.32, P = 0.005), but quite similar in overweight and obese patients (Table 4). In multivariable analysis, the overall adjusted mortality risk did not differ significantly across baseline BMI groups: underweight, HR 1.05, 95% CI 0.59–1.88, P = 0.863; overweight, HR 0.92, 95% CI 0.59–1.45, P = 0.726; obese, HR 0.65, 95% CI 0.32–1.29, P = 0.218. We divided patients into those randomized to enalapril (with or without betablocker) and those randomized to placebo (with or without betablocker). Comparing overweight and overweight/obese patients, respectively, with normalweight patients, the HRs for mortality differed significantly
and borderline-significantly, respectively, between the placebo and enalapril groups: In the placebo group, overweight and overweight/obese patients showed HRs above 1 (increased risk) compared to normal-weight patients, whereas in the enalapril group, overweight and overweight/obese patients showed HRs below 1 (decreased risk) compared to normal-weight patients (Table 4). Thus, there was significant impact of randomized treatment on the association between BMI and mortality (Table 4). The same pattern was seen for obese patients, although with a non-significant P-value for interaction (Table 4). When dividing patients into those with and without betablocker, compared to normal-weight patients, overweight, obese and overweight/obese patients combined had HRs for mortality above 1 (i.e. increased risk) in the non-betablocker group, but HRs below 1 (i.e. decreased risk) in the betablocker group. Test for interaction showed statistical borderline-significance or non-significance with regard to differences between HRs (Table 4). Categorizing patients into those with (angiotensin-converting enzyme inhibitor and/or betablocker) and without (neither angiotensin-converting enzyme inhibitor nor betablocker) neurohormonal blockade, compared to normal-
Table 4 CONSENSUS II: Results from univariable Cox proportional hazard analyses and test for interaction, with BMI group as explanatory variable and normal-weight as reference Overweight (n = 2572)
All patients NH blockade No NH blockade Interaction NH blockade yes/no Enalapril Placebo Interaction enalapril/placebo Betablocker No betablocker Interaction betablocker yes/no
Obese (n = 712)
Overweight/obese (n = 3284)
HR
95% CI
P-value
HR
95% CI
P-value
HR
95% CI
P-value
1.01 0.79 1.38 0.57 0.75 1.41 0.53 0.57 1.10 0.52
0.77–1.33 0.55–1.14 0.90–2.13 0.33–1.01 0.51–1.10 0.93–2.13 0.30–0.94 0.26–1.27 0.82–1.47 0.22–1.23
0.932 0.215 0.14 0.054 0.140 0.105 0.028 0.170 0.546 0.138
0.982 0.90 1.11 0.81 0.92 1.06 0.86 0.30 1.13 0.27
0.66–1.46 0.55–1.49 0.58–2.14 0.36–1.84 0.55–1.52 0.57–1.99 0.38–1.92 0.07–1.33 0.75–1.72 0.06–1.26
0.927 0.685 0.746 0.612 0.732 0.849 0.707 0.113 0.553 0.095
1.00 0.82 1.33 0.62 0.79 1.34 0.59 0.51 1.10 0.46
0.77–1.31 0.58–1.16 0.87–2.02 0.36–1.06 0.55–1.13 0.89–2.01 0.34–1.01 0.23–1.09 0.83–1.46 0.20–1.05
0.969 0.254 0.183 0.079 0.191 0.157 0.053 0.082 0.497 0.065
NH blockade, neurohormonal blockade, i.e. an angiotensin-converting enzyme inhibitor and/or a betablocker.
L.M.A. Kennedy et al. / International Journal of Cardiology 119 (2007) 33–40
37
Fig. 1. CONSENSUS II study. Mortality hazard ratios for baseline body mass index group versus the corresponding normal-weight group, depending on treatment with neurohormonal blockade. Bar graphs show patients receiving any neurohormonal blockade (white) and those not receiving any neurohormonal blockade (dark gray), patients randomized to enalapril treatment (white with black horizontal bars) and those randomized to placebo (black with white horizontal bars), patients with beta-blockers (white with black dots) and patients without beta-blockers (black with white dots). Hazard ratios are for comparison with the corresponding normal-weight group, e.g. overweight patients receiving enalapril are compared with normal-weight patients receiving enalapril. P-values are given for interaction between treatment and no treatment.
weight patients, overweight patients and overweight/obese patients combined had HRs for mortality above 1 (i.e. increased risk) in the non-neurohormonal group, but HRs below 1 (i.e. decreased risk) in the neurohormonal group. Test for interaction showed borderline statistical significance with regard to differences between HRs (Table 4). The same pattern was seen for obese patients, although with a non-significant P-value for interaction (Table 4, Fig. 1). 3.2. 4S The use of neurohormonal blockade was somewhat more common among overweight and obese patients and increased during the study period (Table 3). Only 97 patients received an angiotensin-converting enzyme inhibitor at baseline, whereas 513 patients did at some time point during
the study. In univariable analysis including all patients, compared to normal-weight patients, the overall mortality risk was similar among underweight (HR 1.34, 95% CI 0.96–1.87, P = 0.09), overweight (HR 0.96, 95% CI 0.77– 1.19, P = 0.697) and obese patients (HR 1.22, 95% CI 0.89– 1.68, P = 0.217). In multivariable analysis, the overall adjusted mortality risk differed significantly between underweight and normal-weight patients (HR 1.41, 95% CI 1.01–1.99, P = 0.044) but did not differ significantly between normal-weight and any of the other baseline BMI groups (Table 5). Compared to normal-weight patients, overweight patients and overweight/obese patients combined had HRs for mortality above 1 (i.e. increased risk) among those who did not receive any neurohormonal blockade, but HRs below 1 (i.e. decreased risk) among those receiving neurohormonal
Table 5 4S: Results from multivariable Cox proportional hazard analyses and test for interaction, with BMI group as explanatory variable and normal-weight as reference Overweight (n = 2060)
All patients Neurohormonal blockade No neurohormonal blockade Interaction NH blockade yes/no ACEi No ACEi Interaction ACEi yes/no Betablocker No betablocker Interaction betablocker yes/no
Obese (n = 469)
Overweight/obese (n = 2529)
HR
95% CI
P-value
HR
95% CI
P-value
HR
95% CI
P-value
1.00 0.91 1.11 0.82 0.72 1.05 0.68 1.06 1.00 1.07
0.80–1.25 0.66–1.26 0.81–1.51 0.53–1.29 0.40–1.29 0.83–1.34 0.36–1.29 0.74–1.53 0.75–1.33 0.67–1.69
0.994 0.574 0.519 0.394 0.268 0.671 0.235 0.735 0.982 0.780
1.32 0.97 1.86 0.52 0.68 1.44 0.47 1.15 1.57 0.73
0.96–1.83 0.61–1.56 1.20–2.87 0.28–0.99 0.29–1.61 1.02–2.04 0.19–1.19 0.69–1.91 1.04–2.36 0.38–1.41
0.092 0.913 0.005 0.047 0.382 0.038 0.111 0.602 0.032 0.351
1.06 0.92 1.22 0.75 0.71 1.12 0.63 1.08 1.09 0.99
0.85–1.31 0.68–1.25 0.91–1.64 0.49–1.15 0.40–1.23 0.89–1.41 0.35–1.15 0.76–1.52 0.83–1.43 0.64–1.53
0.624 0.608 0.175 0.191 0.220 0.342 0.134 0.673 0.537 0.964
Betablocker and angiotensin-converting enzyme inhibitor use was assessed by time dependent analysis. NH blockade, neurohormonal blockade, i.e. an angiotensin-converting enzyme inhibitor and/or a betablocker.
38
L.M.A. Kennedy et al. / International Journal of Cardiology 119 (2007) 33–40
blockade (Table 5). However, test for interaction did not show that the HRs differed significantly (Table 5). Obese patients, compared to normal-weight patients, had substantially and significantly increased mortality risk among those who did not receive any neurohormonal blockade, but HR below 1 (decreased mortality risk) among patients who were given such treatment (Table 5). Test for interaction showed that the HR for mortality among obese patients without neurohormonal blockade (versus normal-weight patients without neurohormonal blockade) was significantly higher compared to the HR for mortality among obese patients with neurohormonal blockade (versus normal-weight patients with neurohormonal blockade) (Table 5). When dividing patients into patients with and those without an angiotensin-converting enzyme inhibitor, compared to normal-weight patients, overweight and overweight/ obese patients without an angiotensin-converting enzyme inhibitor had HRs above 1 (increased risk), whereas those with an angiotensin-converting enzyme inhibitor had HRs below 1 (decreased risk) (Table 5). However, test for interaction did not show that the HRs differed significantly (Table 5). Obese patients without an angiotensin-converting enzyme inhibitor had significantly increased mortality risk compared to normal-weight patients without an angiotensin-converting enzyme inhibitor, whereas obese patients with an angiotensin-converting enzyme inhibitor had HR below 1 compared to normal-weight patients with an angiotensin-converting enzyme inhibitor (Table 5). However, test for interaction did not show that the HRs differed significantly (Table 5).
When dividing patients into those with and without a betablocker, compared to normal-weight patients, obese patients without a betablocker had significantly increased mortality risk compared to normal-weight patients without a betablocker, whereas obese patients with a betablocker had HR slightly above 1 compared to normal-weight patients with a betablocker (Table 5). However, test for interaction did not show that the HRs differed significantly (Table 5). Both among patients with and those without a betablocker, overweight and overweight/obese patients combined had HRs at or just above 1, as compared with normal-weight patients (Fig. 2). 4. Discussion The present study in patients following acute myocardial infarction or with stable coronary artery disease shows that, during both long- and short-term follow-up, the mortality risk associated with overweight and obesity was related to treatment with neurohormonal blockers, i.e. angiotensinconverting enzyme inhibitors and/or betablockers. Thus, among patients receiving neurohormonal blockade, overweight/obesity was usually associated with HRs below 1 (i.e. decreased risk) for mortality, compared to normal-weight (with neurohormonal blockers), whereas, among those who did not receive such treatment, overweight/obesity was associated with HRs above 1 (i.e. increased risk) for mortality compared to normal-weight (without neurohormonal blockers). Importantly, some tests for interaction indicated that the association between overweight/obesity and survival
Fig. 2. 4S study. Mortality hazard ratios versus the corresponding normal-weight group, depending on treatment with neurohormonal blockade and baseline body mass index group. Bar graphs show patients receiving any neurohormonal blockade (white) and those not receiving any neurohormonal blockade (dark gray), patients receiving angiotensin-converting enzyme inhibitor treatment (white with black horizontal bars) and those not receiving such treatment (black with white horizontal bars), patients with betablockers (white with black dots) and patients without betablockers (black with white dots). Hazard ratios are for comparison with the corresponding normal-weight group, e.g. overweight patients receiving neurohormonal blockade are compared with normal-weight patients receiving neurohormonal blockade. P-values are given for interaction between treatment and no treatment.
L.M.A. Kennedy et al. / International Journal of Cardiology 119 (2007) 33–40
was reversed by treatment with neurohormonal blockers. In CONSENSUS II this was seen among overweight patients with regard to randomization to enalapril or placebo, and in 4S it was shown among obese patients with regard to treatment with any neurohormonal blockade or not. In addition, in CONSENSUS II, HRs for overweight and overweight/ obese patients, versus the corresponding normal-weight group, differed borderline-significantly between those with and those without treatment with any neurohormonal blockade. The results from prior studies on the prognostic importance of BMI in patients with established coronary artery disease and after acute myocardial infarction are inconsistent [3–16,19]. In a study in survivors of a first acute myocardial infarction, obesity was associated with increased risk of recurrent coronary events during 3 years of follow-up [3]. Another study in patients after acute myocardial infarction showed that obesity was an independent predictor of inhospital death in older, but not in younger patients [4]. On the other hand, 1-year mortality for patients discharged from hospital was significantly less in obese than in normalweight patients, although this was not shown to be independent of age [4]. In a population-based cohort study, postacute myocardial infarction patients who were overweight or obese did not show worse survival, compared to those who were normal-weight [5]. In another study in patients after acute myocardial infarction, there was an insignificant trend towards a positive graded relation between BMI and mortality risk [6]. In the Physicians' Health Study, around 5000 men with a history of acute myocardial infarction or stroke, but without cancer, were followed for 5 years with regard to all-cause and cardiovascular death. A BMI of 28 or greater was not associated with increased risk [12]. In an analysis of patients with stable cardiovascular disease included in the HOPE study, patients in the third tertile of waist circumference had an independently increased risk of acute myocardial infarction, chronic heart failure and total mortality during 4.5 years of follow-up, compared with those in the first tertile [13]. In patients with angiographically verified coronary atherosclerosis, elevated BMI was associated with an increased risk of acute coronary syndrome [11]. In an analysis of patients following acute myocardial infarction included in the TRACE study, high BMI was not associated with increased mortality risk, whereas high waist–hip ratio was, however only in men [14]. In a prior study of patients with complicated acute myocardial infarction, we found that overweight or obesity were not associated with increased mortality risk, compared to normal-weight [19]. Studies on patients undergoing percutaneous coronary intervention have found a protective effect of obesity on outcome [7–9]. In patients undergoing coronary by-pass surgery, those with low BMI were at higher risk for adverse outcome, compared to obese patients [10]. In another study, obesity was not associated with increased in-hospital or 1-year mortality following coronary by-pass surgery [16]. In yet another study of patients after coronary by-pass surgery, those who were overweight or obese had signifi-
39
cantly better survival free from major cardiac or cerebrovascular events, compared to those with normal BMI [15]. Although high BMI is associated with increased risk of developing chronic heart failure [22], higher BMI was associated with a trend toward improved survival in patients with advanced chronic heart failure [23,24]. As shown in the present paper, the impact on the future prognosis of being overweight or obese may well be quite different in patients treated with neurohormonal blockers following an acute myocardial infarction, compared to those not receiving such treatment. Therefore, we suggest that variations in the use of neurohormonal blockers might also possibly explain different study results concerning the impact of high BMI in patients with cardiovascular disease [3–16,19]. The rationale for such an explanation is the increasing evidence of the importance of neurohormonal activation to the pathophysiology of obesity-related cardiovascular disease. There is evidence of existence of a local renin–angiotensin system in adipose tissue [17,18], which may play an important role in the pathophysiology of obesity and obesityrelated cardiovascular disorders. This is supported by the found positive correlation between BMI in human subjects and plasma angiotensin concentrations, renin activity and angiotensin-converting enzyme activity [18], indicating that obesity is associated with systemic activation of the renin– angiotensin system [25]. In this respect the possible interference of angiotensin II with the insulin-signaling cascade is also of potential importance to the development of diabetes and cardiovascular disease [17]. Increased sympathetic activity in obesity has also been described, suggesting that neurohormonal activation may be of importance to the pathophysiology of obesity-related hypertension and cardiovascular disease [18]. Hence, if the renin–angiotensin system and/or the sympathetic nervous system are blocked in overweight and obese patients following acute myocardial infarction, any beneficial effects associated with overweight/ obesity might come into play, such as a higher metabolic reserve, which has been shown to provide better resistance against the metabolic stress associated with chronic heart failure [24]. Our data does not offer any definitive explanation to the finding that the impact of neurohormonal treatment on the prognostic importance of BMI was statistically significant among overweight patients in CONSENSUS II but among obese patients in 4S. Differences in the time of follow-up, somewhat different patient populations and/or differences in treatment may be of relevance. In CONSENSUS II, the angiotensin-converting enzyme inhibitor enalapril constituted most of the neurohormonal blockade, whereas betablockade was the principal neurohormonal blockade in 4S. Patients in CONSENSUS II were all post-acute myocardial infarction patients, whereas those in 4S had stable coronary artery disease, although most had a history of a prior acute myocardial infarction. The time of follow-up was much longer in 4S compared to CONSENSUS II. Furthermore,
40
L.M.A. Kennedy et al. / International Journal of Cardiology 119 (2007) 33–40
tests for statistical interaction are inherently low powered and consequently there is a significant risk of type 2 error, whereby a real interaction may be undetected. A limitation to this study is obviously the retrospective nature of the analysis. In addition, in 4S patients were not randomized to treatment with neurohormonal blockers. However, in CONSENSUS II patients were actually randomized to treatment with enalapril or placebo. Another limitation is that only height and weight, but neither waist circumference nor waist–hip ratio were measured. Waist circumference reflects abdominal obesity and might therefore be the most desired measurement. However, since BMI is generally used to define overweight/obesity in clinical practice and in guidelines, analyses based on BMI are of relevance.
[6]
[7]
[8]
[9]
[10]
[11]
5. Conclusion In patients after acute myocardial infarction or with stable coronary artery disease, overweight and obesity was in general not associated with increased mortality risk compared to normal-weight. However, among those not receiving treatment with angiotensin-converting enzyme inhibitors and/or betablockers, the mortality risk for overweight and obese patients was generally increased compared to normal-weight patients, whereas it was usually decreased for overweight and obese patients versus normal-weight patients among those receiving angiotensin-converting enzyme inhibitors and/or betablockers. Some tests for interaction indicated that neurohormonal blockade may attenuate or possibly even invert the relationship between high BMI and mortality. Neurohormonal blockade may thus partly explain the so-called obesity paradox. This study is hypothesis generating and future investigation should look further into this possible relationship.
[12]
[13]
[14]
[15]
[16]
[17]
[18]
Acknowledgement [19]
This study was supported by a grant from Merck and Co. Linn Kennedy received a research grant from the Swedish Heart–Lung Foundation.
[20]
References [1] Eckel RH. Obesity and heart disease: a statement for healthcare professionals from the Nutrition Committee, American Heart Association. Circulation 1997;96:3248–50. [2] Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath Jr CW. Bodymass index and mortality in a prospective cohort of U.S. adults. N Engl J Med 1999;341:1097–105. [3] Rea TD, Heckbert SR, Kaplan RC, et al. Body mass index and the risk of recurrent coronary events following acute myocardial infarction. Am J Cardiol 2001;88:467–72. [4] Hoit BD, Gilpin EA, Maisel AA, Henning H, Carlisle J, Ross Jr J. Influence of obesity on morbidity and mortality after acute myocardial infarction. Am Heart J 1987;114:1334–41. [5] Lopez-Jimenez F, Jacobsen SJ, Reeder GS, Weston SA, Meverden RA, Roger VL. Prevalence and secular trends of excess body weight and
[21]
[22] [23]
[24]
[25]
impact on outcomes after myocardial infarction in the community. Chest 2004;125:1205–12. Rana JS, Mukamal KJ, Morgan JP, Muller JE, Mittleman MA. Obesity and the risk of death after acute myocardial infarction. Am Heart J 2004;147:841–6. Gruberg L, Weissman NJ, Waksman R, et al. The impact of obesity on the short-term and long-term outcomes after percutaneous coronary intervention: the obesity paradox? J Am Coll Cardiol 2002;39:578–84. Gurm HS, Brennan DM, Booth J, Tcheng JE, Lincoff AM, Topol EJ. Impact of body mass index on outcome after percutaneous coronary intervention (the obesity paradox). Am J Cardiol 2002;90:42–5. Powell BD, Lennon RJ, Lerman A, et al. Association of body mass index with outcome after percutaneous coronary intervention. Am J Cardiol 2003;91:472–6. Potapov EV, Loebe M, Anker S, et al. Impact of body mass index on outcome in patients after coronary artery bypass grafting with and without valve surgery. Eur Heart J 2003;24:1933–41. Wolk R, Berger P, Lennon RJ, Brilakis ES, Somers VK. Body mass index: a risk factor for unstable angina and myocardial infarction in patients with angiographically confirmed coronary artery disease. Circulation 2003;108:2206–11. Widlansky ME, Sesso HD, Rexrode KM, Manson JE, Gaziano JM. Body mass index and total and cardiovascular mortality in men with a history of cardiovascular disease. Arch Intern Med 2004;164: 2326–32. Dagenais GR, Yi Q, Mann JF, Bosch J, Pogue J, Yusuf S. Prognostic impact of body weight and abdominal obesity in women and men with cardiovascular disease. Am Heart J 2005;149:54–60. Kragelund C, Hassager C, Hildebrandt P, Torp-Pedersen C, Kober L. Impact of obesity on long-term prognosis following acute myocardial infarction. Int J Cardiol 2005;98:123–31. Gruberg L, Mercado N, Milo S, et al. Impact of body mass index on the outcome of patients with multivessel disease randomized to either coronary artery bypass grafting or stenting in the ARTS trial: The obesity paradox II? Am J Cardiol 2005;95:439–44. Lindhout AH, Wouters CW, Noyez L. Influence of obesity on inhospital and early mortality and morbidity after myocardial revascularization. Eur J Cardiothorac Surg 2004;26:535–41. Goossens GH, Blaak EE, van Baak MA. Possible involvement of the adipose tissue renin–angiotensin system in the pathophysiology of obesity and obesity-related disorders. Obes Rev 2003;4:43–55. Engeli S, Sharma AM. Role of adipose tissue for cardiovascular–renal regulation in health and disease. Horm Metab Res 2000;32:485–99. Kennedy LM, Dickstein K, Anker SD, Kristianson K, Willenheimer R. The prognostic importance of body mass index after complicated myocardial infarction. J Am Coll Cardiol 2005;45:156–8. Swedberg K, Held P, Kjekshus J, Rasmussen K, Ryden L, Wedel H. Effects of the early administration of enalapril on mortality in patients with acute myocardial infarction. Results of the Cooperative New Scandinavian Enalapril Survival Study II (CONSENSUS II). N Engl J Med 1992;327:678–84. Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S). Lancet 1994;344:1383–9. Kenchaiah S, Evans JC, Levy D, et al. Obesity and the risk of heart failure. N Engl J Med 2002;347:305–13. Horwich TB, Fonarow GC, Hamilton MA, MacLellan WR, Woo MA, Tillisch JH. The relationship between obesity and mortality in patients with heart failure. J Am Coll Cardiol 2001;38:789–95. Davos CH, Doehner W, Rauchhaus M, et al. Body mass and survival in patients with chronic heart failure without cachexia: the importance of obesity. J Card Fail 2003;9:29–35. Barton M, Carmona R, Ortmann J, Krieger JE, Traupe T. Obesityassociated activation of angiotensin and endothelin in the cardiovascular system. Int J Biochem Cell Biol 2003;35:826–37.
Impact of neurohormonal blockade on association between body mass index and mortality Linn M.A. Kennedy a,⁎, Stefan D. Anker b,c , John Kjekshus d , Tom J. Cook e , Ronnie Willenheimer a a
c
Department of Cardiology, Malmö University Hospital, Lund University, Malmö, Sweden b Department of Clinical Cardiology, National Heart and Lung Institute, London, UK Division of Applied Cachexia Research, Department of Cardiology, Charité Campus Virchow Klinikum, Berlin, Germany d Department of Cardiology, Stavanger University Hospital, Oslo, Norway e Merck and Co., Whitehouse Station, NJ, USA Received 13 February 2006; accepted 19 June 2006 Available online 17 October 2006
Abstract Background: The prognostic impact of body mass index (BMI) in patients following acute myocardial infarction (AMI) may be altered by neurohormonal blockade. Methods: The impact of neurohormonal blockade on the association between BMI and mortality was examined in 5548 patients following AMI (CONSENSUS II), 50% receiving enalapril and 7% beta-blockade, and in 4367 patients with coronary artery disease (CAD) (4S), 79% with prior AMI, 12% receiving ACEi and 67% beta-blockade. Median follow-up was 0.4 and 5.2 years, respectively. Patients were categorized into 4 BMI groups: Underweight, b 22.00; normal-weight, 22.00–24.99; overweight, 25.00–29.99; obese, ≥ 30.00 kg/m2. Multivariable analysis adjusted for demographics, patient history, physical examination, biochemistry and medication. Results: CONSENSUS II: Overall, adjusted mortality (n = 301) risk was similar across BMI groups. Comparing overweight with normalweight patients, the hazard ratios (HRs) for mortality differed significantly (P = 0.028) between patients randomized to placebo (HR 1.41) and enalapril (HR 0.75). 4S: Overall, adjusted mortality (n = 421) risk was similar for normal-weight, overweight and obese patients. In a time-dependent analysis for drug use, comparing obese with normal-weight patients, the HRs for mortality differed significantly (P = 0.047) between patients without (HR 1.86) and those with (HR 0.97) neurohormonal blockade. Conclusion: In patients after AMI or with CAD, high BMI was associated with increased mortality risk among patients not receiving neurohormonal blockade, but with decreased or neutral mortality risk among those receiving neurohormonal blockade. Tests for interaction indicate that neurohormonal blockade may attenuate the relationship between high BMI and increased mortality risk. Neurohormonal blockade may thus partly explain the so-called obesity paradox. © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Acute myocardial infarction; Coronary artery disease; Body mass index; Neurohormonal blockade; Mortality
1. Introduction Obesity is a serious independent risk factor for coronary artery disease and premature death, and is related to cardiovascular risk factors such as hypertension, dyslipidemia and insulin resistance [1,2]. However, the prognostic importance ⁎ Corresponding author. Tel.: +46 40 331000; fax: +46 40 336209. E-mail address: [email protected] (L.M.A. Kennedy). 0167-5273/$ - see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ijcard.2006.06.064
of overweight and obesity in patients with coronary artery disease and following acute myocardial infarction is not yet fully understood. Studies have shown diverse and even contradictory results; high body mass index (BMI) has had a negative, neutral or even positive effect on survival [3–16]. In patients with coronary artery disease undergoing percutaneous coronary intervention, overweight and obesity have been associated with a better prognosis, a finding sometimes referred to as “the obesity paradox”[7].
34
L.M.A. Kennedy et al. / International Journal of Cardiology 119 (2007) 33–40
There is evidence that local adipose tissue and systemic renin–angiotensin system activation, as well as obesityrelated increased sympathetic activity may play an important role in the pathophysiology of obesity-related cardiovascular disorders [17,18]. Hence, it may be of benefit to block the renin–angiotensin system and the sympathetic nervous system in overweight and obese patients. In a previous analysis of the OPTIMAAL trial we found that, in 5477 patients with complicated acute myocardial infarction, overweight and obese patients had similar adjusted mortality risk compared to normal-weight patients, during a follow-up of 2.8 years [19]. These findings might possibly be due to the effects of renin– angiotensin system inhibition by angiotensin-converting enzyme inhibitors or angiotensin-receptor blockers, and/or inhibition of the sympathetic nervous system by betablockade; all patients received an angiotensin-converting enzyme inhibitor or an angiotensin-receptor blockers and 75% were given beta-blockade in the OPTIMAAL trial [19]. Therefore, we hypothesized that overweight/obesity is associated with increased mortality risk in patients not receiving neurohormonal blockade – i.e. angiotensinconverting enzyme inhibitors, angiotensin-receptor blockers or betablockers – but not in patients receiving such treatment. In order to test our hypothesis we evaluated the impact of treatment with neurohormonal blockade on the relationship between BMI and prognosis in patients after acute myocardial infarction or with stable coronary artery disease, included in the CONSENSUS II [20] and 4S trials [21]. 2. Methods 2.1. Patients and trial design The 6090 patients in the CONSENSUS II trial had acute myocardial infarction and were randomized to double-blind treatment with enalapril or placebo [20]. Beta-blockers were optional. The primary endpoint was all-cause mortality. The 4444 patients included in the 4S trial had angina pectoris or previous acute myocardial infarction (79%) and received randomized double-blind treatment with simvastatin or placebo [21]. Angiotensin-converting enzyme inhibitors and betablockers were optional. The primary endpoint was allcause mortality.
(22.00–24.99 kg/m2), overweight (25.00–29.99 kg/m2), obese (N/= 30 kg/m2). In multivariable analyses, adjustment was made for several baseline variables comprising demographics, patient history, medication, physical examination and biochemical analyses. 2.3. Statistics The value of baseline BMI group to predict mortality was tested in uni- and multivariable Cox Proportional Hazard models. In a second step, we tested the impact of treatment with neurohormonal blockade on the association between baseline BMI group and mortality. Due to statistical power issues, these analyses were performed with the overweight and obese groups separately as well as combined. Formal tests for interaction depending on treatment with neurohormonal blockade were performed. In CONSENSUS II, in-hospital use of betablockers and angiotensin-converting enzyme inhibitors was considered use of that treatment. Due to few events in the analyses when dividing the material, test for interaction was only performed in univariable model in CONSENSUS II. In 4S treatment with betablockers and angiotensin-converting enzyme inhibitors was used as a time dependent variable. In this analysis, patients were considered to be on treatment if they had been so for at least 14 days, and to be off treatment if they had not received any such treatment within 14 days. Test for interaction was performed in a multivariable model. Baseline covariables were screened for possible inclusion in a multivariable model. Hazard ratios (HR) and 95% confidence intervals (CI) were estimated for the effect of BMI group relative to the normal BMI group. Covariables found to be significantly (P b 0.05) associated with mortality in univariable analysis were tested for internal correlation, using Spearman rank correlation test for categorical variables and linear regression analysis for continuous variables. In all analyses, R N 0.3 denoted significant internal correlation, and only the variable with the greatest association with mortality was qualified for further analysis. Variables remaining after this procedure were included in the corresponding multivariable model. The covariable selection method resulted in the inclusion into the CONSENSUS II multivariable model of baseline loop diuretic use, systolic blood pressure, heart rate, age, creati-
2.2. Design of the present analysis
Table 1 Baseline BMI by baseline BMI group
In post-hoc analysis we assessed the impact of treatment with an angiotensin-converting enzyme inhibitor and/or a betablocker on the association between baseline BMI group and mortality (no patient received an angiotensin-receptor blockers in CONSENSUS II or 4S). Prior to data analysis, patients were categorized into 4 groups according to baseline BMI: underweight (b22.00 kg/m2), normal-weight
BMI group
Underweight Normal Overweight Obese BMI, body mass index.
CONSENSUS II
4S
N
%
N
%
640 1624 2572 712
11.5 29.3 46.4 12.8
364 1474 2060 469
8.3 33.8 47.2 10.7
L.M.A. Kennedy et al. / International Journal of Cardiology 119 (2007) 33–40
35
Table 2a CONSENSUS II: Baseline characteristics by baseline BMI group Baseline characteristic
Baseline BMI group Underweight
Normal
N = 640
Age (years) Pulse rate (beats/min) Potassium (mmol/L) Sodium (mmol/L) Serum creatinine (mmol/L) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg)
Female gender History of AMI History of hypertension History of diabetes History of CHF Baseline statin use Baseline aspirin use Baseline beta blocker use Baseline diuretic use Baseline warfarin use
Overweight
N = 1624
Obese
N = 2572
N = 712
Mean
S.D.
Mean
S.D.
Mean
S.D.
Mean
S.D.
67.93 76.51 4.05 139.42 94.28 130.73 77.37
11.34 16.58 0.44 3.19 25.28 19.50 11.97
66.64 74.14 4.04 139.60 95.70 131.93 79.24
11.04 14.58 0.44 3.19 23.04 19.13 11.89
64.36 74.27 4.04 139.67 96.82 133.78 80.49
11.01 15.00 0.43 3.09 22.47 19.98 12.14
63.49 77.58 4.06 139.65 95.76 138.09 82.91
11.29 16.60 0.44 3.18 26.58 21.30 12.64
n
%
n
%
n
%
n
%
247 141 128 57 41 4 11 83 267 61
38.59 22.03 20.00 8.91 6.41 0.63 1.72 12.97 41.72 9.53
378 381 351 141 95 17 30 230 614 171
23.28 23.46 21.61 8.68 5.85 1.05 1.85 14.16 37.81 10.53
590 591 726 286 121 15 35 399 962 244
22.94 22.98 28.23 11.12 4.70 0.58 1.36 15.51 37.40 9.49
257 186 299 133 76 6 9 120 340 42
36.10 26.12 41.99 18.68 10.67 0.84 1.24 16.85 47.75 5.90
nine, potassium, sodium and total lactate dehydrogenase. In 4S, the multivariable model covariable selection resulted in the inclusion into the model of age, gender, smoking status, hemoglobin, history of diabetes, hypertension and claudication, randomized treatment with simvastatin/placebo, and inclusion into the 4S trial due to acute myocardial infarction or not.
3. Results In both studies, around half of the patients were overweight, one quarter were normal-weight and a little more than 10% were obese (Table 1). Baseline characteristics of the patients included in the respective trials are seen in Table 2a (CONSENSUS II) and Table 2b (4S). In
Table 2b 4S: Baseline characteristics by baseline BMI group Baseline characteristic
Age (years) Glucose (mmol/L) HDL (mmol/L) Hemoglobin (mmol/L) LDL cholesterol (mmol/L) Pulse rate (beats/min) Serum creatinine (mmol/L) Systolic blood pressure (mm Hg) Triglycerides (mmol/L)
Female gender Current smoker History of diabetes History of AMI Baseline aspirin Baseline beta blocker Baseline diuretic Baseline ACEi and/or beta blocker
Baseline BMI group Underweight
Normal
Overweight
Obese
N = 364
N = 1474
N = 2060
N = 469
Mean
S.D.
59.2 99.5 1.35 85.7 4.77 64.5 87.1 135.4 1.26
7.0 20.3 0.34 65.7 0.67 10.7 19.8 20.0 0.43
Mean
n
%
n
%
118 131 9 296 123 181 11 185
32.4 36.0 2.5 81.3 33.8 49.7 3.0 50.8
262 414 41 1179 570 776 81 793
17.8 28.1 2.8 80.0 38.7 52.6 5.5 53.8
59.0 102.3 1.21 86.1 4.94 64.0 90.6 136.5 1.41
S.D. 7.0 21.2 0.30 67.9 0.67 10.1 21.2 19.3 0.50
Mean
S.D.
Mean
S.D.
7.12 26.98 0.28 0.69 0.65 9.92 15.06 19.17 0.50
57.86 111.44 1.13 9.34 4.85 64.84 92.24 144.20 1.69
7.26 30.13 0.25 0.72 0.64 9.71 15.03 20.43 0.50
n
%
n
%
308 457 109 1625 747 1237 122 1260
14.95 22.18 5.29 78.88 36.26 60.05 5.92 61.17
122 98 42 373 164 311 61 318
26.01 20.90 8.96 79.53 34.97 66.31 13.01 67.80
58.33 106.22 1.15 9.28 4.86 63.36 93.47 139.81 1.56
36
L.M.A. Kennedy et al. / International Journal of Cardiology 119 (2007) 33–40
Table 3 CONSENSUS II: Neurohormonal blockade by baseline BMI group Neurohormonal blockade
ACEi Beta-blocker ACEi and betablocker Neither ACEi nor betablocker
Underweight
Normal
Overweight
Obese
n
%
n
%
n
%
n
%
295 43 40 262
46.09 6.72 6.25 40.94
697 114 116 697
42.92 7.02 7.14 42.92
1036 202 197 1137
40.28 7.85 7.66 44.21
317 56 64 275
44.52 7.87 8.99 38.62
4S: Neurohormonal blockade use at least 14 days during study by baseline BMI group Neurohormonal blockade
ACEi Betablocker ACEi and/or betablocker
Underweight
Normal
Overweight
Obese
n
%
n
%
n
%
n
%
28 209 220
7.7 57.4 60.4
148 918 970
10.0 62.3 65.8
252 1432 1501
12.2 69.5 72.9
85 353 371
18.12 75.27 79.10
ACEi, angiotensin-converting enzyme inhibitor.
CONSENSUS II, the follow-up period had a median length of 0.4 years (139 days, range 2–181 days) and in 4S it was 5.2 years (range 11–2267 days). During follow-up, there were 301 deaths in CONSENSUS II and 421 deaths in 4S. 3.1. CONSENSUS II The use of neurohormonal blockade was similar across the baseline BMI groups (Table 3). In univariable analysis including all patients, compared to normal-weight patients, the overall mortality risk was significantly higher in underweight patients (HR 1.64, 95% CI 1.16–2.32, P = 0.005), but quite similar in overweight and obese patients (Table 4). In multivariable analysis, the overall adjusted mortality risk did not differ significantly across baseline BMI groups: underweight, HR 1.05, 95% CI 0.59–1.88, P = 0.863; overweight, HR 0.92, 95% CI 0.59–1.45, P = 0.726; obese, HR 0.65, 95% CI 0.32–1.29, P = 0.218. We divided patients into those randomized to enalapril (with or without betablocker) and those randomized to placebo (with or without betablocker). Comparing overweight and overweight/obese patients, respectively, with normalweight patients, the HRs for mortality differed significantly
and borderline-significantly, respectively, between the placebo and enalapril groups: In the placebo group, overweight and overweight/obese patients showed HRs above 1 (increased risk) compared to normal-weight patients, whereas in the enalapril group, overweight and overweight/obese patients showed HRs below 1 (decreased risk) compared to normal-weight patients (Table 4). Thus, there was significant impact of randomized treatment on the association between BMI and mortality (Table 4). The same pattern was seen for obese patients, although with a non-significant P-value for interaction (Table 4). When dividing patients into those with and without betablocker, compared to normal-weight patients, overweight, obese and overweight/obese patients combined had HRs for mortality above 1 (i.e. increased risk) in the non-betablocker group, but HRs below 1 (i.e. decreased risk) in the betablocker group. Test for interaction showed statistical borderline-significance or non-significance with regard to differences between HRs (Table 4). Categorizing patients into those with (angiotensin-converting enzyme inhibitor and/or betablocker) and without (neither angiotensin-converting enzyme inhibitor nor betablocker) neurohormonal blockade, compared to normal-
Table 4 CONSENSUS II: Results from univariable Cox proportional hazard analyses and test for interaction, with BMI group as explanatory variable and normal-weight as reference Overweight (n = 2572)
All patients NH blockade No NH blockade Interaction NH blockade yes/no Enalapril Placebo Interaction enalapril/placebo Betablocker No betablocker Interaction betablocker yes/no
Obese (n = 712)
Overweight/obese (n = 3284)
HR
95% CI
P-value
HR
95% CI
P-value
HR
95% CI
P-value
1.01 0.79 1.38 0.57 0.75 1.41 0.53 0.57 1.10 0.52
0.77–1.33 0.55–1.14 0.90–2.13 0.33–1.01 0.51–1.10 0.93–2.13 0.30–0.94 0.26–1.27 0.82–1.47 0.22–1.23
0.932 0.215 0.14 0.054 0.140 0.105 0.028 0.170 0.546 0.138
0.982 0.90 1.11 0.81 0.92 1.06 0.86 0.30 1.13 0.27
0.66–1.46 0.55–1.49 0.58–2.14 0.36–1.84 0.55–1.52 0.57–1.99 0.38–1.92 0.07–1.33 0.75–1.72 0.06–1.26
0.927 0.685 0.746 0.612 0.732 0.849 0.707 0.113 0.553 0.095
1.00 0.82 1.33 0.62 0.79 1.34 0.59 0.51 1.10 0.46
0.77–1.31 0.58–1.16 0.87–2.02 0.36–1.06 0.55–1.13 0.89–2.01 0.34–1.01 0.23–1.09 0.83–1.46 0.20–1.05
0.969 0.254 0.183 0.079 0.191 0.157 0.053 0.082 0.497 0.065
NH blockade, neurohormonal blockade, i.e. an angiotensin-converting enzyme inhibitor and/or a betablocker.
L.M.A. Kennedy et al. / International Journal of Cardiology 119 (2007) 33–40
37
Fig. 1. CONSENSUS II study. Mortality hazard ratios for baseline body mass index group versus the corresponding normal-weight group, depending on treatment with neurohormonal blockade. Bar graphs show patients receiving any neurohormonal blockade (white) and those not receiving any neurohormonal blockade (dark gray), patients randomized to enalapril treatment (white with black horizontal bars) and those randomized to placebo (black with white horizontal bars), patients with beta-blockers (white with black dots) and patients without beta-blockers (black with white dots). Hazard ratios are for comparison with the corresponding normal-weight group, e.g. overweight patients receiving enalapril are compared with normal-weight patients receiving enalapril. P-values are given for interaction between treatment and no treatment.
weight patients, overweight patients and overweight/obese patients combined had HRs for mortality above 1 (i.e. increased risk) in the non-neurohormonal group, but HRs below 1 (i.e. decreased risk) in the neurohormonal group. Test for interaction showed borderline statistical significance with regard to differences between HRs (Table 4). The same pattern was seen for obese patients, although with a non-significant P-value for interaction (Table 4, Fig. 1). 3.2. 4S The use of neurohormonal blockade was somewhat more common among overweight and obese patients and increased during the study period (Table 3). Only 97 patients received an angiotensin-converting enzyme inhibitor at baseline, whereas 513 patients did at some time point during
the study. In univariable analysis including all patients, compared to normal-weight patients, the overall mortality risk was similar among underweight (HR 1.34, 95% CI 0.96–1.87, P = 0.09), overweight (HR 0.96, 95% CI 0.77– 1.19, P = 0.697) and obese patients (HR 1.22, 95% CI 0.89– 1.68, P = 0.217). In multivariable analysis, the overall adjusted mortality risk differed significantly between underweight and normal-weight patients (HR 1.41, 95% CI 1.01–1.99, P = 0.044) but did not differ significantly between normal-weight and any of the other baseline BMI groups (Table 5). Compared to normal-weight patients, overweight patients and overweight/obese patients combined had HRs for mortality above 1 (i.e. increased risk) among those who did not receive any neurohormonal blockade, but HRs below 1 (i.e. decreased risk) among those receiving neurohormonal
Table 5 4S: Results from multivariable Cox proportional hazard analyses and test for interaction, with BMI group as explanatory variable and normal-weight as reference Overweight (n = 2060)
All patients Neurohormonal blockade No neurohormonal blockade Interaction NH blockade yes/no ACEi No ACEi Interaction ACEi yes/no Betablocker No betablocker Interaction betablocker yes/no
Obese (n = 469)
Overweight/obese (n = 2529)
HR
95% CI
P-value
HR
95% CI
P-value
HR
95% CI
P-value
1.00 0.91 1.11 0.82 0.72 1.05 0.68 1.06 1.00 1.07
0.80–1.25 0.66–1.26 0.81–1.51 0.53–1.29 0.40–1.29 0.83–1.34 0.36–1.29 0.74–1.53 0.75–1.33 0.67–1.69
0.994 0.574 0.519 0.394 0.268 0.671 0.235 0.735 0.982 0.780
1.32 0.97 1.86 0.52 0.68 1.44 0.47 1.15 1.57 0.73
0.96–1.83 0.61–1.56 1.20–2.87 0.28–0.99 0.29–1.61 1.02–2.04 0.19–1.19 0.69–1.91 1.04–2.36 0.38–1.41
0.092 0.913 0.005 0.047 0.382 0.038 0.111 0.602 0.032 0.351
1.06 0.92 1.22 0.75 0.71 1.12 0.63 1.08 1.09 0.99
0.85–1.31 0.68–1.25 0.91–1.64 0.49–1.15 0.40–1.23 0.89–1.41 0.35–1.15 0.76–1.52 0.83–1.43 0.64–1.53
0.624 0.608 0.175 0.191 0.220 0.342 0.134 0.673 0.537 0.964
Betablocker and angiotensin-converting enzyme inhibitor use was assessed by time dependent analysis. NH blockade, neurohormonal blockade, i.e. an angiotensin-converting enzyme inhibitor and/or a betablocker.
38
L.M.A. Kennedy et al. / International Journal of Cardiology 119 (2007) 33–40
blockade (Table 5). However, test for interaction did not show that the HRs differed significantly (Table 5). Obese patients, compared to normal-weight patients, had substantially and significantly increased mortality risk among those who did not receive any neurohormonal blockade, but HR below 1 (decreased mortality risk) among patients who were given such treatment (Table 5). Test for interaction showed that the HR for mortality among obese patients without neurohormonal blockade (versus normal-weight patients without neurohormonal blockade) was significantly higher compared to the HR for mortality among obese patients with neurohormonal blockade (versus normal-weight patients with neurohormonal blockade) (Table 5). When dividing patients into patients with and those without an angiotensin-converting enzyme inhibitor, compared to normal-weight patients, overweight and overweight/ obese patients without an angiotensin-converting enzyme inhibitor had HRs above 1 (increased risk), whereas those with an angiotensin-converting enzyme inhibitor had HRs below 1 (decreased risk) (Table 5). However, test for interaction did not show that the HRs differed significantly (Table 5). Obese patients without an angiotensin-converting enzyme inhibitor had significantly increased mortality risk compared to normal-weight patients without an angiotensin-converting enzyme inhibitor, whereas obese patients with an angiotensin-converting enzyme inhibitor had HR below 1 compared to normal-weight patients with an angiotensin-converting enzyme inhibitor (Table 5). However, test for interaction did not show that the HRs differed significantly (Table 5).
When dividing patients into those with and without a betablocker, compared to normal-weight patients, obese patients without a betablocker had significantly increased mortality risk compared to normal-weight patients without a betablocker, whereas obese patients with a betablocker had HR slightly above 1 compared to normal-weight patients with a betablocker (Table 5). However, test for interaction did not show that the HRs differed significantly (Table 5). Both among patients with and those without a betablocker, overweight and overweight/obese patients combined had HRs at or just above 1, as compared with normal-weight patients (Fig. 2). 4. Discussion The present study in patients following acute myocardial infarction or with stable coronary artery disease shows that, during both long- and short-term follow-up, the mortality risk associated with overweight and obesity was related to treatment with neurohormonal blockers, i.e. angiotensinconverting enzyme inhibitors and/or betablockers. Thus, among patients receiving neurohormonal blockade, overweight/obesity was usually associated with HRs below 1 (i.e. decreased risk) for mortality, compared to normal-weight (with neurohormonal blockers), whereas, among those who did not receive such treatment, overweight/obesity was associated with HRs above 1 (i.e. increased risk) for mortality compared to normal-weight (without neurohormonal blockers). Importantly, some tests for interaction indicated that the association between overweight/obesity and survival
Fig. 2. 4S study. Mortality hazard ratios versus the corresponding normal-weight group, depending on treatment with neurohormonal blockade and baseline body mass index group. Bar graphs show patients receiving any neurohormonal blockade (white) and those not receiving any neurohormonal blockade (dark gray), patients receiving angiotensin-converting enzyme inhibitor treatment (white with black horizontal bars) and those not receiving such treatment (black with white horizontal bars), patients with betablockers (white with black dots) and patients without betablockers (black with white dots). Hazard ratios are for comparison with the corresponding normal-weight group, e.g. overweight patients receiving neurohormonal blockade are compared with normal-weight patients receiving neurohormonal blockade. P-values are given for interaction between treatment and no treatment.
L.M.A. Kennedy et al. / International Journal of Cardiology 119 (2007) 33–40
was reversed by treatment with neurohormonal blockers. In CONSENSUS II this was seen among overweight patients with regard to randomization to enalapril or placebo, and in 4S it was shown among obese patients with regard to treatment with any neurohormonal blockade or not. In addition, in CONSENSUS II, HRs for overweight and overweight/ obese patients, versus the corresponding normal-weight group, differed borderline-significantly between those with and those without treatment with any neurohormonal blockade. The results from prior studies on the prognostic importance of BMI in patients with established coronary artery disease and after acute myocardial infarction are inconsistent [3–16,19]. In a study in survivors of a first acute myocardial infarction, obesity was associated with increased risk of recurrent coronary events during 3 years of follow-up [3]. Another study in patients after acute myocardial infarction showed that obesity was an independent predictor of inhospital death in older, but not in younger patients [4]. On the other hand, 1-year mortality for patients discharged from hospital was significantly less in obese than in normalweight patients, although this was not shown to be independent of age [4]. In a population-based cohort study, postacute myocardial infarction patients who were overweight or obese did not show worse survival, compared to those who were normal-weight [5]. In another study in patients after acute myocardial infarction, there was an insignificant trend towards a positive graded relation between BMI and mortality risk [6]. In the Physicians' Health Study, around 5000 men with a history of acute myocardial infarction or stroke, but without cancer, were followed for 5 years with regard to all-cause and cardiovascular death. A BMI of 28 or greater was not associated with increased risk [12]. In an analysis of patients with stable cardiovascular disease included in the HOPE study, patients in the third tertile of waist circumference had an independently increased risk of acute myocardial infarction, chronic heart failure and total mortality during 4.5 years of follow-up, compared with those in the first tertile [13]. In patients with angiographically verified coronary atherosclerosis, elevated BMI was associated with an increased risk of acute coronary syndrome [11]. In an analysis of patients following acute myocardial infarction included in the TRACE study, high BMI was not associated with increased mortality risk, whereas high waist–hip ratio was, however only in men [14]. In a prior study of patients with complicated acute myocardial infarction, we found that overweight or obesity were not associated with increased mortality risk, compared to normal-weight [19]. Studies on patients undergoing percutaneous coronary intervention have found a protective effect of obesity on outcome [7–9]. In patients undergoing coronary by-pass surgery, those with low BMI were at higher risk for adverse outcome, compared to obese patients [10]. In another study, obesity was not associated with increased in-hospital or 1-year mortality following coronary by-pass surgery [16]. In yet another study of patients after coronary by-pass surgery, those who were overweight or obese had signifi-
39
cantly better survival free from major cardiac or cerebrovascular events, compared to those with normal BMI [15]. Although high BMI is associated with increased risk of developing chronic heart failure [22], higher BMI was associated with a trend toward improved survival in patients with advanced chronic heart failure [23,24]. As shown in the present paper, the impact on the future prognosis of being overweight or obese may well be quite different in patients treated with neurohormonal blockers following an acute myocardial infarction, compared to those not receiving such treatment. Therefore, we suggest that variations in the use of neurohormonal blockers might also possibly explain different study results concerning the impact of high BMI in patients with cardiovascular disease [3–16,19]. The rationale for such an explanation is the increasing evidence of the importance of neurohormonal activation to the pathophysiology of obesity-related cardiovascular disease. There is evidence of existence of a local renin–angiotensin system in adipose tissue [17,18], which may play an important role in the pathophysiology of obesity and obesityrelated cardiovascular disorders. This is supported by the found positive correlation between BMI in human subjects and plasma angiotensin concentrations, renin activity and angiotensin-converting enzyme activity [18], indicating that obesity is associated with systemic activation of the renin– angiotensin system [25]. In this respect the possible interference of angiotensin II with the insulin-signaling cascade is also of potential importance to the development of diabetes and cardiovascular disease [17]. Increased sympathetic activity in obesity has also been described, suggesting that neurohormonal activation may be of importance to the pathophysiology of obesity-related hypertension and cardiovascular disease [18]. Hence, if the renin–angiotensin system and/or the sympathetic nervous system are blocked in overweight and obese patients following acute myocardial infarction, any beneficial effects associated with overweight/ obesity might come into play, such as a higher metabolic reserve, which has been shown to provide better resistance against the metabolic stress associated with chronic heart failure [24]. Our data does not offer any definitive explanation to the finding that the impact of neurohormonal treatment on the prognostic importance of BMI was statistically significant among overweight patients in CONSENSUS II but among obese patients in 4S. Differences in the time of follow-up, somewhat different patient populations and/or differences in treatment may be of relevance. In CONSENSUS II, the angiotensin-converting enzyme inhibitor enalapril constituted most of the neurohormonal blockade, whereas betablockade was the principal neurohormonal blockade in 4S. Patients in CONSENSUS II were all post-acute myocardial infarction patients, whereas those in 4S had stable coronary artery disease, although most had a history of a prior acute myocardial infarction. The time of follow-up was much longer in 4S compared to CONSENSUS II. Furthermore,
40
L.M.A. Kennedy et al. / International Journal of Cardiology 119 (2007) 33–40
tests for statistical interaction are inherently low powered and consequently there is a significant risk of type 2 error, whereby a real interaction may be undetected. A limitation to this study is obviously the retrospective nature of the analysis. In addition, in 4S patients were not randomized to treatment with neurohormonal blockers. However, in CONSENSUS II patients were actually randomized to treatment with enalapril or placebo. Another limitation is that only height and weight, but neither waist circumference nor waist–hip ratio were measured. Waist circumference reflects abdominal obesity and might therefore be the most desired measurement. However, since BMI is generally used to define overweight/obesity in clinical practice and in guidelines, analyses based on BMI are of relevance.
[6]
[7]
[8]
[9]
[10]
[11]
5. Conclusion In patients after acute myocardial infarction or with stable coronary artery disease, overweight and obesity was in general not associated with increased mortality risk compared to normal-weight. However, among those not receiving treatment with angiotensin-converting enzyme inhibitors and/or betablockers, the mortality risk for overweight and obese patients was generally increased compared to normal-weight patients, whereas it was usually decreased for overweight and obese patients versus normal-weight patients among those receiving angiotensin-converting enzyme inhibitors and/or betablockers. Some tests for interaction indicated that neurohormonal blockade may attenuate or possibly even invert the relationship between high BMI and mortality. Neurohormonal blockade may thus partly explain the so-called obesity paradox. This study is hypothesis generating and future investigation should look further into this possible relationship.
[12]
[13]
[14]
[15]
[16]
[17]
[18]
Acknowledgement [19]
This study was supported by a grant from Merck and Co. Linn Kennedy received a research grant from the Swedish Heart–Lung Foundation.
[20]
References [1] Eckel RH. Obesity and heart disease: a statement for healthcare professionals from the Nutrition Committee, American Heart Association. Circulation 1997;96:3248–50. [2] Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath Jr CW. Bodymass index and mortality in a prospective cohort of U.S. adults. N Engl J Med 1999;341:1097–105. [3] Rea TD, Heckbert SR, Kaplan RC, et al. Body mass index and the risk of recurrent coronary events following acute myocardial infarction. Am J Cardiol 2001;88:467–72. [4] Hoit BD, Gilpin EA, Maisel AA, Henning H, Carlisle J, Ross Jr J. Influence of obesity on morbidity and mortality after acute myocardial infarction. Am Heart J 1987;114:1334–41. [5] Lopez-Jimenez F, Jacobsen SJ, Reeder GS, Weston SA, Meverden RA, Roger VL. Prevalence and secular trends of excess body weight and
[21]
[22] [23]
[24]
[25]
impact on outcomes after myocardial infarction in the community. Chest 2004;125:1205–12. Rana JS, Mukamal KJ, Morgan JP, Muller JE, Mittleman MA. Obesity and the risk of death after acute myocardial infarction. Am Heart J 2004;147:841–6. Gruberg L, Weissman NJ, Waksman R, et al. The impact of obesity on the short-term and long-term outcomes after percutaneous coronary intervention: the obesity paradox? J Am Coll Cardiol 2002;39:578–84. Gurm HS, Brennan DM, Booth J, Tcheng JE, Lincoff AM, Topol EJ. Impact of body mass index on outcome after percutaneous coronary intervention (the obesity paradox). Am J Cardiol 2002;90:42–5. Powell BD, Lennon RJ, Lerman A, et al. Association of body mass index with outcome after percutaneous coronary intervention. Am J Cardiol 2003;91:472–6. Potapov EV, Loebe M, Anker S, et al. Impact of body mass index on outcome in patients after coronary artery bypass grafting with and without valve surgery. Eur Heart J 2003;24:1933–41. Wolk R, Berger P, Lennon RJ, Brilakis ES, Somers VK. Body mass index: a risk factor for unstable angina and myocardial infarction in patients with angiographically confirmed coronary artery disease. Circulation 2003;108:2206–11. Widlansky ME, Sesso HD, Rexrode KM, Manson JE, Gaziano JM. Body mass index and total and cardiovascular mortality in men with a history of cardiovascular disease. Arch Intern Med 2004;164: 2326–32. Dagenais GR, Yi Q, Mann JF, Bosch J, Pogue J, Yusuf S. Prognostic impact of body weight and abdominal obesity in women and men with cardiovascular disease. Am Heart J 2005;149:54–60. Kragelund C, Hassager C, Hildebrandt P, Torp-Pedersen C, Kober L. Impact of obesity on long-term prognosis following acute myocardial infarction. Int J Cardiol 2005;98:123–31. Gruberg L, Mercado N, Milo S, et al. Impact of body mass index on the outcome of patients with multivessel disease randomized to either coronary artery bypass grafting or stenting in the ARTS trial: The obesity paradox II? Am J Cardiol 2005;95:439–44. Lindhout AH, Wouters CW, Noyez L. Influence of obesity on inhospital and early mortality and morbidity after myocardial revascularization. Eur J Cardiothorac Surg 2004;26:535–41. Goossens GH, Blaak EE, van Baak MA. Possible involvement of the adipose tissue renin–angiotensin system in the pathophysiology of obesity and obesity-related disorders. Obes Rev 2003;4:43–55. Engeli S, Sharma AM. Role of adipose tissue for cardiovascular–renal regulation in health and disease. Horm Metab Res 2000;32:485–99. Kennedy LM, Dickstein K, Anker SD, Kristianson K, Willenheimer R. The prognostic importance of body mass index after complicated myocardial infarction. J Am Coll Cardiol 2005;45:156–8. Swedberg K, Held P, Kjekshus J, Rasmussen K, Ryden L, Wedel H. Effects of the early administration of enalapril on mortality in patients with acute myocardial infarction. Results of the Cooperative New Scandinavian Enalapril Survival Study II (CONSENSUS II). N Engl J Med 1992;327:678–84. Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S). Lancet 1994;344:1383–9. Kenchaiah S, Evans JC, Levy D, et al. Obesity and the risk of heart failure. N Engl J Med 2002;347:305–13. Horwich TB, Fonarow GC, Hamilton MA, MacLellan WR, Woo MA, Tillisch JH. The relationship between obesity and mortality in patients with heart failure. J Am Coll Cardiol 2001;38:789–95. Davos CH, Doehner W, Rauchhaus M, et al. Body mass and survival in patients with chronic heart failure without cachexia: the importance of obesity. J Card Fail 2003;9:29–35. Barton M, Carmona R, Ortmann J, Krieger JE, Traupe T. Obesityassociated activation of angiotensin and endothelin in the cardiovascular system. Int J Biochem Cell Biol 2003;35:826–37.